Testing a systems research organizing model for behavioral health

Abstract

The purpose of this multifaceted research project was to explain and predict mental health outcomes through testing of a systems research organizing model using pre-existing behavioral health consumer-oriented data. Community Partnership of Southern Arizona provided the setting for the participation of its members in the statewide 2001 Mental Health Statistics Improvement Program Consumer Perception Survey. The sample for this study consisted of 214 adult member survey respondents. The Systems Research Organizing Model for Behavioral Health (SROM-BH) provided the conceptual framework for examining client risk adjustment characteristics and cost and access factors that interact with consumer participation processes to affect consumer perception of quality and health related quality of life. The American Academy of Nursing's Quality Health Outcomes Model and The University of Arizona Nursing Systems Core's System Research Organizing Model informed the development of the SROM-BH that extends this work through its adaptation for use within the context of behavioral health. Composite indices were developed for five model variables, implying that composites or latent variables can be developed from existing data when there is fastidious attention to theory and the conceptual definitions of the constructs. Eight hypothesized positive predictor and three unhypothesized negative predictor relationships were supported. Three hypothesized positive predictor relationships were not supported. Consumer participation in treatment planning, the intervention of interest in this study had an effect (either direct or indirect) on all five outcome variables. Reexamination of model relationships with a larger sample and continued testing of the survey instrument for psychometric performance is recommended. Further model testing using separate scales or methods is needed in order to reduce method effect and to determine the full strength of the findings. Use of structural equation modeling may offer a more precise test of the theoretical framework, strengthen support for instrument subscale construct validity through confirmatory factor analysis, and may provide an opportunity for analysis of recursive paths. Further development of recovery authentication, a concept developed as a result of this study, may contribute to a broadened understanding of opportunities to promote recovery and moderate the loss of self that is associated with mental illness.